1. Field of the Invention
The present invention relates to an image processing apparatus, an image processing method and a computer program.
2. Description of the Related Art
Various kinds of noise are contained in raw data output from a charge coupled device (CCD) or a complementary metal oxide semiconductor (CMOS) device which are being used as imaging devices in digital still cameras.
For example, when an image is developed by converting raw data into a signal in a different format, it is necessary that noise contained in the signal after the conversion is effectively removed. For example, when the raw data are converted into image data (for example, YCC data) including luminance signals and color difference signals, image noise may be classified into luminance noise and color noise. It is well known that image quality can be improved by removing such color noise.
As one of methods for removing noise contained in raw data, for example, there has been proposed a method for removing noise from Gr/R/B/Gb data, as disclosed in Japanese Unexamined Patent Application Publication No. 2005-311455. However, when such noise are removed from the Gr/R/B/Gb data, the data may not be separated into luminance signals and color difference signals, the luminance signals may be affected by such a noise removing process. This may result in inauthentic color where an edge portion in which luminance has been changed is colored or has a different luminance.
As another of method for removing color noise, there has been proposed a method for removing noise from YCC (YCrCb) data including luminance signals and color difference signals which are obtained by subjecting raw data to color separation such as de-mosaic process or the like. This method can remove color noise without having an effect on the luminance signals by performing noise removing process for the color difference signals. However, when a high-level de-mosaic process is performed which considers edge direction in a state where significant noise is contained, wrong determination of an edge may be made due to noise. This may result in the inclusion of inauthentic color or color noise in an image after de-mosaic process. In addition, since the YCC data passes through signal processing such as de-mosaic process or the like, the property of noise contained in the color difference signal Cr/Cb may become complicated so that color noise may not be completely removed. In addition, if noise reduction to remove the color noise is strongly carried out, an edge may be dimmed.
For the purpose of addressing this problem, there has been proposed a method for transforming raw data into color difference signals through simple and easy de-mosaic process without considering edge direction, applying noise removing process to remove color noise from the color difference signals, and retransforming the color difference signals into raw data (see Japanese Unexamined Patent Application Publication No. 2008-153836). In this method, YCC data are produced as a high-level de-mosaic process including process based on edge determination is performed for the retransformed raw data after removing the noise. The method disclosed in Japanese Unexamined Patent Application Publication No. 2008-153836 can prevent the color difference signals from being deteriorated based on a wrong determination of an edge due to noise since simple and easy de-mosaic process without considering edge direction is performed to remove color noise.